Characterization of Recombinant Corn GlutathioneS-Transferase Isoforms I, II, III, and IV

Pesticide Biochemistry and Physiology 63, 127–138 (1999) Article ID pest.1999.2396, available online at http://www.idealibrary.com on

Characterization of Recombinant Corn Glutathione S-Transferase Isoforms I, II, III, and IV Alfred Sommer and Peter Bo¨ger Lehrstuhl fu¨r Physiologie und Biochemie der Pflanzen, Universita¨t Konstanz, D-78457 Konstanz, Germany Received September 1, 1998; accepted January 11, 1999 Glutathione S-transferases (GSTs) are involved in detoxification of a wide variety of electrophilic compounds including herbicides. Several corn isoforms (GSTs) have been studied for their ability to conjugate these substrates with reduced glutathione (GSH). Three cDNAs, encoding corn GST subunits of 29, 27, and 26 kDa, respectively, were cloned into expression systems in Escherichia coli. N-terminal 6xHis-tagged recombinant GST isoforms I, II, III, and IV were purified with nickel-nitrilotriacetic acid (Ni-NTA) metalaffinity chromatography and were analyzed biochemically. As the corn enzymes, each recombinant GST isoform also consists of two subunits. Using three different GST-substrates, recombinant isoforms showed similar substrate specificities as natural corn GSTs. Some GST isoforms may be involved in the defense response to oxidative stress in plants. Besides standard GST activities, inactivation of endogenous, toxic a,b-unsaturated aldehydes was measured. Furthermore two recombinant GST isoforms (GST II and GST IV) showed high glutathione peroxidase activity using three different organic hydroperoxides as substrates. Apparently, GST isoforms including the 27-kDa subunit show glutathione peroxidase activity. q1999 Academic Press

Key Words: conjugation; His-tag affinity purification; isomerization; peroxidase; recombinant GST isoforms; subunits.

INTRODUCTION

Glutathione S-transferases (GSTs,1 EC 2.5.1.18) are a group of enzymes that catalyze the conjugation of a wide range of hydrophobic, electrophilic, usually cytotoxic substrates with the tripeptide glutathione (g-glutamyl-cysteinylglycin, GSH). Polar nontoxic peptide conjugates are accessible to further metabolic steps (1). It becomes clear that conjugation activity is not their only role, and this study will check for 1 Abbreviations and chemical names used: ampr, ampicillin resistance; Bis-Tris propane, 1,3-bis[tris(hydroxymethyl)methylamino]propane; CDNB, 1-chloro-2,4-dinitrobenzene; DTT, dithiothreitol; EDTA, ethylenediaminetetraacetic acid; GSH, reduced glutathione; GST, glutathione S-transferase; kanr, kanamycin resistance; metazachlor, 2-chloro-N-(2,6-dimethylphenyl)-N-(1H-pyrazol-1-ylmethyl)acetamide; NTA, Ni-nitrilotriacetic acid; PAGE, polyacrylamide gel electrophoresis; protox, protoporphyrinogen oxidase; SDS, sodium dodecylsulfate; thiadiazolidine BW78, 5-(4-bromophenylimino)-3,4-tetramethylene-1,3,4thiadiazolidin-2-one; triazolidine BW85, 4-(4-bromophenyl)-1,2-tetramethylene-1,2, 4-triazolidin-3-one-5-thione.

additional activities. Almost all organisms possess multiple GST isoenzymes, which are classified in alpha, mu, pi, sigma, and theta class (1, 2). Plant GSTs were first identified and have been intensively studied because of their ability to detoxify herbicides (1, 3, 4). Plants with higher GST activity levels withstand exposure to herbicides which kill susceptible species (1, 5). Enzymatic activities towards different classes of herbicides like chloroacetamides (e.g., acetochlor, alachlor, metazachlor, or metolachlor), chloro-S-triazines (e.g., atrazine), thiocarbamates (like EPTC sulfoxide) (6), and peroxidizing herbicides are compounds which have been investigated (7). Peroxidizing herbicides like oxyfluorfen or cyclic imides lead to rapid phytotoxic degradation of plant cell constituents (8). In corn, isomerization of thiadiazolidine to peroxidizing triazolidine herbicides is catalyzed by GSTs, resulting in a bioactivation of these herbicides (9). Most cytosolic plant GSTs in crops have been

127 0048-3575/99 $30.00 Copyright q 1999 by Academic Press All rights of reproduction in any form reserved.

128

¨ GER SOMMER AND BO

classified (10, 11). Class type I (theta), type II (zeta), type III (tau), and type IV were derived from phylogenetic DNA analysis. At present, eight cytosolic GST isoenzymes in corn have been described.2 Seven dimeric isoforms are involved in herbicide metabolism. These isoenzymes include five different subunits: GST29 (29 kDa), GST27 (27 kDa), GST26 (26 kDa; three partly different cDNA sequences were determined), all belonging to class theta, Zm GST V (class tau; 28.5 kDa) and Zm GST VI (class tau; 27.5 kDa). Some cDNA clones of corn GST subunits have been isolated (GST29 (12, 13), GST27 (12, 14), GST26 (2, 15, 16), and Zm GST V (17)). GST I is a constitutive (and somewhat safener-inducible) homodimer of two GST29 subunits and shows activity towards alachlor, atrazine, and CDNB (1, 9, 18). This latter isoenzyme also catalyzes the isomerization of peroxidizing thiadiazolidin-one herbicides to the more active triazolidin-one-thiones (9, 19, 20). GST II is a safener-inducible (by flurazole and dichlormid (1)) heterodimer (a GST29 and a GST27 subunit), which shows activity toward alachlor and CDNB (9, 12, 18, 21) and which isomerizes thiadiazolidin-one herbicides much better than GST I (9, 19, 20). GST IV is a safener-inducible (by benoxacor) homodimer of two GST27 subunits with activity toward acetochlor, alachlor, and metolachlor (1, 22). Homodimer GST III (two GST26 subunits) may be induced by cadmium (23) and dichlormid (1) and is active towards alachlor, metolachlor, and CDNB (1, 15, 16, 24). Zm GST IIII is a constitutive heterodimer with activity against chloroacetamides and fluorodifen (25). Homodimer Zm GST V-V is selectively inducible by the safener dichlormid (17) with activity toward diphenyl ether herbicides. Heterodimer Zm GST V-VI is an auxin-inducible isoform with activity against metolachlor (17). BZ 2 is an 2 A revised nomenclature (17, 25) was proposed for these GSTs based on their subunit composition with a prefix indicating the plant source Zea mays (Zm). Corn isoenzyme GST I corresponds to Zm GST I-I, GST II to Zm GST I-II, GST III to Zm GST III-III, and GST IV to Zm GST II-II.

inducible type III corn GST (26 kDa) with activity towards CDNB and other natural substrates (26). Activity of a ninth corn enzyme against unsaturated phenylpropanoids has been attributed to an ascorbate peroxidase enzyme and not to a GST (27). We report the purification of four recombinant N-terminal 6xHis-tagged corn GST isoforms from Escherichia coli and their biochemical characterization. Activities of these enzymes with endogenous, toxic alkenal substrate analogues, as well as GST-mediated peroxidase activity were studied. MATERIALS AND METHODS

cDNA Cloning and Construction of Plasmids Techniques of DNA manipulations were used as described (28). All PCR reactions were performed in a Programmable Thermal Controller PTC-100 (MJ Research Inc., Watertown, MA) using a standard program. The first step was a 5-min denaturation of the cDNAs, followed by 31 cycles with an annealing temperature of 558C for 2 min, 90 s for elongation at 728C, and denaturation at 948C for 1 min. Two steps for annealing (2 min at 558C) and elongation (5 min at 728C) finished reactions. Primers for PCR were synthesized at MWG-BIOTECH (Ebersberg, Germany). Taq-DNA polymerase was from Eurobio (Raunheim, Germany), kits from Qiagen (Hilden, Germany) were used for DNA isolation and purification. All restriction enzymes, alkaline phosphatase (from calf intestine), and T4 DNA ligase were from Boehringer (Mannheim, Germany). Expression vectors pQE30 (ampr ) and pQE31 (ampr ) were from Qiagen (Hilden, Germany). cDNAs of corn GST-subunits GST29 (12, 13), GST27 (12, 14), and GST26 (15), cloned in plasmids (pIJ12), (pIJ21), and (pGTC20), respectively, were used as templates for PCR. Primers for PCR were designed so that the entire coding regions of the cDNAs would be amplified and subsequently expressed in frame. Sequence analysis of DNA and proteins was performed with PC Gene program (Version 9.0, 1993; IntelliGenetics, Geest, Belgium).

RECOMBINANT GST ISOFORMS

Plasmid pGST27. Expression vector of the GST27 cDNA (recombinant protein: GST IV). Primers for PCR, 58-GAG AAC AGG ATC CAT GGC TAC GCC-38 (forward primer) and 58-CGC TTC GCC AGC TTC ATC GTC AC38 (reverse primer) were designed according to the GST27 cDNA sequence as described. BamHI and HindIII restriction sites were introduced at the 58 and 38 ends, respectively. The single DNA fragment (700 bp) was isolated and restricted with BamHI and HindIII. The purified DNA was cloned into the pQE30 expression vector. Plasmid pGST29. Expression vector of the GST29 cDNA (recombinant protein GST I). PCR-primers, 58-GTC TGG CAT GCC ATG GCT CCG-38 (forward primer) and 58-CGT CAT AGA AGC TTC ATC ACT GAA CAA GC-38 (reverse primer) were designed according to the GST29 cDNA sequence as described. SphI and HindIII restriction sites were introduced at the 58 and 38 ends, respectively. The single DNA fragment (740 bp) was isolated and restricted with SphI and HindIII. The purified DNA was cloned into the pQE31 expression vector. Plasmid pGST27-29. Coexpression vector of the GST27 and GST29 cDNA (recombinant protein GST II). This coexpression vector was constructed to yield a single mRNA transcription product which includes the GST27 and GST29 coding sequences. Each GST subunit is expected to be translated separately (mRNA includes two ribosomal binding sites for each coding sequence) resulting in 6xHis-tagged GST subunits GST29 and GST27. Transcription and translation of these subunits in close neighborhood should promote formation of heterodimeric GST II. Primers for PCR, 58-GTG AGC GGA TAA AGC TTT CAC ACA G-38 (forward primer) and 58CGT CAT AGA AGC TTC ATC ACT GAA CAA GC-38 (reverse primer), were designed according to the the pGST29 DNA sequence (construction see above). Two HindIII restriction sites were introduced at the 58 and 38 ends, respectively. The single DNA fragment (780 bp) was isolated and restricted with BamHI and HindIII. pGST27 was restricted with HindIII and

129

treated with alkaline phosphatase. The purified product from PCR was cloned into the prepared pGST27 vector. Plasmid pGST26. Expression vector of the GST26 cDNA (recombinant protein GST III). Primers for PCR, 58-GCA GCA GGA TCC ATG GCG C-38 (forward primer) and 58-GGC AAC GCA AGC TTA GGT CAA GC-38 (reverse primer) were designed according to the GST26 cDNA sequence as described. BamHI and HindIII restriction sites were introduced at the 58 and 38 ends, respectively. The single DNA fragment (700 bp) was isolated and restricted with BamHI and HindIII. The purified DNA was cloned into the pQE30 expression vector. Each plasmid was transformed by electroporation into E. coli M15 host strain (from Qiagen, Hilden, Germany) containing the repressor plasmid pREP4 (kanr). Expression of cDNAs and Protein Purification under Native Conditions Transformed E. coli strains were grown at 378C in 2 liters LB medium containing 100 mg/ ml ampicillin and 25 mg/ml kanamycin. At OD600 5 0.7–0.9, cells were induced with 1 mM isopropyl-b-D-thiogalactopyranoside (IPTG) and cultivated at 28–328C overnight. The culture was cooled on ice for 30 min. The cells were harvested by centrifugation (48C, 4,000g, 20 min), resuspended in 5 vol of lysis buffer (50 mM K-phosphate, pH 7.5, containing 300 mM NaCl, 10 mM imidazole, 10% (v/v) glycerol, and 0.1% (v/v) Tween 20) and ruptured in a French pressure cell (three times at 29.7 MPa). The lysate was stirred on ice for 30 min and centrifuged at 48C for 20 min (25,000g). Protein was isolated in a batch purification. The homogenate was incubated with nickel-nitrilotriacetic acid (Ni-NTA) agarose resin (Qiagen, Hilden, Germany), under slight shaking for 1 h on ice. The protein–resin complex was collected by a small column and washed with lysis buffer and washing buffer (lysis buffer containing 20 mM imidazole). Elution of 6xHis-tagged protein was performed with a gradient of washing and elution buffer (lysis buffer containing 500 mM

130

SOMMER AND BÖGER

imidazole). Fractions of 2 ml were collected and analyzed by SDS±PAGE. Electrophoretically homogeneous protein (GST I, GST IV, and GST III) was desalted into storage buffer (50 mM Kphosphate, pH 7.5, containing 1 mM EDTA, 10% (v/v) glycerol, and 0.1% (v/v) Tween 20) using PD-10 columns (Pharmacia, Freiburg, Germany). Purification was also performed with K-phosphate or Na-phosphate buffers without Tween 20. Protein from coexpression of GST27 and GST29 cDNA was desalted with PD-10 columns into buffer A (20 mM Bis-Tris propane, pH 7.4, containing 1 mM EDTA, 1 mM MgCl2, 1 mM DTT, and 10% (v/v) glycerol) or buffer A containing 0.1% (v/v) Tween 20. Separation of three possible isoforms from coexpression was performed at room temperature on Mono Q column (HR 10/10 from Pharmacia, Freiburg, Germany) using buffer A with a 0 to 500 mM NaCl gradient. Fractions (2 ml) were collected and analyzed by SDS±PAGE. Electrophoretically homogeneous GST II protein was desalted into storage buffer (with or without Tween 20) using PD-10 columns. Purified proteins were stored at 2208C until used for further experiments. Polyacrylamide Gel Electrophoresis (PAGE) Expression of the transformed E. coli strains as well as purity, molecular weight, and quantitative amounts of recombiant GST isoforms were analyzed by 15% SDS±PAGE according to Laemmli. Native PAGE was performed at 48C to analyze the stability of recombinat GST isoforms under native conditions. The 7.5%-gel contained no denaturating SDS, but an additive of 10% (v/v) glycerol. Molecular Weight Calculation (Gel Filtration) Native apparent molecular weights (MW) of recombinant GST isoforms were determined by gel filtration, which was performed on Superose 12 (HR 10/30; Pharmacia, Freiburg, Germany) using a buffer containing 50 mM K-phosphate, pH 7.5, 150 mM Na-chloride and 10% (v/v)

glycerol. The column was calibrated with ribonuclease A (MW 13.7 kDa), chymotrypsinogen A (MW 25 kDa), ovalbumin (MW 43 kDa), bovine serum albumin (MW 67 kDa), alcohol dehydrogenase (from yeast, MW 150 kDa), bamylase (from sweet potato, MW 200 kDa), and blue dextran (MW 2000 kDa). Flow rate was 20 ml/h and protein elution was monitored for A280. Protein Content The concentration of protein was determined by the method of Bradford (29) using bovine serum albumin as protein standard. Glutathione S-Transferase Assays Conjugation of CDNB (1-chloro-2,4-dinitrobenzene; Sigma, Deisenhofen, Germany) and reduced glutathione (GSH) was measured spectrophotometrically (9, 30) at 340 nm (e 340 5 9.6 mM21 cm21) and 308C. The assay (1 ml) contained 100 mM K-phosphate, pH 7.0, 2 mM CDNB, 2 mM GSH, and one recombinant GST (in K-phosphate buffer containing Tween 20; protein contents of GST I 0.2 mg, GST II 0.4 mg, GST IV 10 mg, or GST III 5 mg). Values for nonenzymatic conjugation were substracted (0.52 mmol/h 3 ml assay). Turnover of 14C-labeled metazachlor (2chloro-N-(2,6-dimethylphenyl)-N-(1H-pyrazol1-yl-methyl)acetamide) was performed as described (9) at 358C. The assay (0.2 ml) contained 100 mM K-phosphate, pH 7.0, 5 mM GSH, 0.25 mM metazachlor, and one recombinant GST (in K-phosphate buffer without Tween 20; protein contents of GST I 12 mg, GST II 1 mg, GST IV 1 mg, or GST III 10 mg). Values for nonenzymatic conjugation were substracted (0.085 mmol/h 3 ml assay). Isomerization of thiadiazolidine BW78 (5-(4bromophenylimino)-3,4-tetramethylene-1,3,4thiadiazolidin-2-one) was carried out as described at 308C (9). The assay contained 100 mM K-phosphate, pH 6.8, 1 mM GSH, 0.1 mM thiadiazolidine, and one recombinant GST (in K-phosphate buffer without Tween 20; protein contents of GST I 5 mg, GST II 0.2 mg, GST IV 0.2 mg, or GST III 12.5 mg). No nonenzymatic

RECOMBINANT GST ISOFORMS

isomerization to triazolidine BW85 (4-(4-bromophenyl)-1,2-tetramethylene-1,2,4-triazolidine-3-one-5-thione) was detected during reaction time. Conversion of ethacrynic acid (Sigma, Deisenhofen, Germany) was measured at 270 nm (31) (ε270 5 5 mM21 cm21) and 308C. The assay (1 ml) contained 50 mM K-phosphate, pH 6.5, 5 mM GSH, 0.2 mM ethacrynic acid, and one recombinant GST (in K-phosphate buffer containing Tween 20; protein contents of GST I 20 mg, GST II 20 mg, GST IV 20 mg, or GST III 20 mg). Values for nonenzymatic conjugation were substracted (0.73 mmol/h 3 ml assay). Conjugation of crotonaldehyde (Fluka, Deisenhofen, Germany) was measured at 230 nm (32) (ε230 5 10.7 mM21 cm21) and 308C. The assay (1 ml) contained 50 mM K-phosphate, pH 6.5, 1 mM GSH, 0.1 mM crotonaldehyde, and one recombinant GST (in K-phosphate buffer containing Tween 20; protein contents of GST I 20 mg, GST II 20 mg, GST IV 20 mg, or GST III 20 mg). Values for nonenzymatic conversion were substracted (0.12 mmol/h 3 ml assay). Glutathione Peroxidase Assays Glutathione peroxidase activity (33, 34) was measured using a glutathione reductase coupled assay to monitor the oxidation of GSH. Standard glutathione peroxidase assay (1 ml) was performed in 50 mM K-phosphate buffer, pH 7.0, containing 2 mM GSH, 0.2 mM NADPH (Sigma, Deisenhofen, Germany), 2 mM EDTA, 0.5 units glutathione reductase, and an organic hydroperoxide substrate (five independent experiments for every substrate). Consumption of NADPH at 308C was monitored spectrophotometrically at 340 nm. Conversion of cumene hydroperoxide (Sigma, Deisenhofen, Germany) (35) was determined using 2 mM hydroperoxide and one recombinant GST (in K-phosphate buffer containing Tween 20; protein contents of GST I 50 mg, GST II 1.5 mg, GST IV 1 mg, or GST III 50 mg). Values for nonenzymatic oxidation of NADPH were substracted (0.55 mmol/h 3 ml assay).

131

The rate with tert.-butyl hydroperoxide (Sigma, Deisenhofen, Germany) (34) was measured in standard assay containing 2 mM of hydroperoxide and one recombinant GST (in Kphosphate buffer containing Tween 20; protein contents of GST I 50 mg, GST II 30 mg, GST IV 20 mg, or GST III 50 mg). Values for nonenzymatic oxidation of NADPH were substracted (0.66 mmol/h 3 ml assay). Conversion of linolenic acid hydroperoxide (36) was determined using 0.15 mM hydroperoxide and one recombinant GST (in K-phosphate buffer containing Tween 20; protein contents of GST I 20 mg, GST II 0.6 mg, GST IV 0.8 mg, or GST III 10 mg). Values for nonenzymatic oxidation of NADPH were substracted (0.2 mmol/h 3 ml assay). Preparation of Linolenic Acid Hydroperoxide Linolenic acid (9,12,15-octadecatrienoic acid) was converted enzymatically (37) to linolenic acid hydroperoxide (13-hydroperoxy9,11,15-octadecatrienoic acid) using soybean lipoxygenase (EC 1.13.11.12; 64,000 units/mg; Sigma, Deisenhofen, Germany). Hydroperoxide was purified over 3-ml C18 reversed-phase sample preparation columns (from Waters, Eschborn, Germany) and eluted with methyl formate. The eluate was evaporated to dryness under nitrogen gas, resolved in absolute ethanol, and stored at 2708C until used for further experiments. Content of hydroperoxide was measured in standard glutathione peroxidase assay using recombinant GST IV as peroxidase. Total consumption of NADPH was determined from small samples of hydroperoxide solution applied to the assay. Values for nonenzymatic decrease of NADPH were substracted. Chemicals and Statistics All chemicals used were of highest purity available. [Phenyl-U-14C]metazachlor (11,6 mCi/mmol 5 429 MBq/mmol) was a generous gift from BASF AG (Limburgerhof, Germany). Thiadiazolidine BW78 and triazolidine BW85 were from Y. Sato and K. Wakabayashi (Department of Agricultural Chemistry, Tamagawa

¨ GER SOMMER AND BO

132

University, Tokyo, Japan). Unless indicated otherwise the results documented in the tables represent mean values with maximum standard deviation from three independent experiments. RESULTS AND DISCUSSION

Purification of Recombinant GST Isoenzymes Expression of GST cDNAs was performed at 28–308C (instead at 378C) to minimize protease activity and formation of inclusion bodies. Addition of detergent (e.g., Tween 20) is required for concentrated protein solutions. Stability of recombinant corn GSTs depends on isoform (stability: I 5 III . II . IV). Expression of GST29, GST27, and GST26 cDNA resulted in accumulation of the corresponding recombinant 63His-tagged GST subunits. Most recombinant protein was soluble and could be isolated in a batch purification with Ni-NTA agarose resin to electrophoretic homogeneity (Fig. 1). The coexpression vector of the GST27 and

GST29 cDNA (plasmid pGST27-29) was constructed to yield a single mRNA transcription product which contains the two GST coding sequences. Each GST subunit is translated separately by a common mRNA, because each coding sequence has its own ribosomal binding site. The result of this coexpression system are 6 3 His-tagged GST subunits GST29 and GST27. Transcription and translation of these subunits in close neighborhood should promote formation of heterodimer GST II. Since no details of formation of heterodimeric GST II in corn was known, it was not clear whether this coexpression vector would lead to formation of recombinant GST II in E. coli. Three possible isoenzymes (GST I, GST II, and GST IV) were expected from coexpression. Surprisingly GST29 and GST27 subunits were isolated in a ratio of 1:3 (Fig. 1). The mixture of recombinant GST isoenzymes was separated on FPLC Mono Q column using a 0–500 mM NaCl gradient (data not shown), and the eluted fractions were analyzed by SDS–PAGE. Small amounts of the

FIG. 1. SDS–PAGE analysis of GST cDNA expression in E. coli and purification of recombinant corn GST isoforms. (A) Isoforms containing GST29 or GST27 subunits; M, marker proteins with apparent molecular weights; 1, E. coli control; 2, expression of GST27 cDNA in E. coli; 3, purified 6xHis-tagged protein GST IV; 4, expression of GST29 cDNA; 5, purified 6xHis-tagged GST I; 6, coexpression of GST29 and GST27 cDNA; 7, purified 6xHis-tagged protein mixture from coexpression; 8, separated GST II heterodimer. (B) Isoform containing GST26 subunit; M, marker proteins; 1, E. coli control; 2, expression of GST26 cDNA in E. coli; 3, purified 6xHis-tagged protein GST III.

133

RECOMBINANT GST ISOFORMS

homodimer GST I eluted at 75–100 mM NaCl, much larger quantities of heterodimer GST II and homodimer GST IV eluted at 125–175 mM and 175–225 mM NaCl, respectively. The elution profile of recombinant isoforms (GST I and GST II) corresponded to natural corn GSTs (9). Seemingly, formation of heterodimer GST II is favored, because almost no homodimer GST I was found and formation of homodimer GST IV is caused by excess expression of subunit GST27. All recombinant GST subunits carry a N-terminal 6xHis-tag resulting in a slight increase of the molecular weights (GST29, 12.0 kDa; GST27 and GST26, 11.6 kDa). Some molecular characteristics of recombinant GST subunits derived from cDNA sequence are shown in Table 1. Calculated isoelectric points correspond to measured values of natural corn GST subunits (GST29: pI 5 6.06 (15); GST27: pI 5 6.1 (14); GST26: pI 5 6.34 (15)). Each recombinant GST isoenzyme was stable as dimeric form under native conditions. Native PAGE of recombinant isoforms containing GST29 or GST27 subunits showed a single band per isoform with different distances from starting point (Fig. 2). Molecular weights of native isoforms were calculated from linear fit of protein standards performed by gel filtration. Each recombinant GST (molecular weights: GST I, 51 6 2 kDa; GST II, 54 6 4 kDa; GST IV, 56 6 3 kDa; and GST III, 48 6 2 kDa) eluted as a single peak and the determined values of molecular weights were as estimated for dimeric forms (Fig. 3).

FIG. 2. Native PAGE of recombinant corn GST isoforms containing GST29 or GST27 subunits. 1, native GST IV; 2, native GST II; 3, native GST I.

Substrate Specificities of Recombinant GST Isoenzymes Comparing purified recombinant GST isoforms, their activities differed markedly (Table 2). Using CDNB as substrate, GST I showed high activitiy, GST II less, GST III low, and GST IV nearly no activity. Apparently, GST isoforms, both homo- and heterodimers including subunit GST29 show CDNB conjugation activity. Conjugation of metazachlor was performed with high rates by GST II or GST IV and low ones by GST I or GST III. No data for metazachlor activity of natural GST IV and GST III are available in literature. Apparently, GST isoforms including subunit GST27 exhibit metazachlor conversion activity. GST I showed low, and GST III nearly no isomerization of thiadiazolidine BW78 to its isomer triazolidine BW85. High isomerization rates were determined in assays with GST II and GST IV. GST isoforms with a GST27 subunit are

TABLE 1 Molecular Characteristics of Recombinant Corn GST Subunits GST29, GST27, and GST26 Recombinant corn GST subunit GST29 GST27 GST26

N-terminal 6xHis-tag region

Number of amino acids

Calculated mol. wt. (kDa)

Calculated isoelectric point

Met-Arg-Gly-Ser-6xHis-Thr-Asp-Pro-His-AlaMet-Arg-Gly-Ser-6xHis-Gly-SerMet-Arg-Gly-Ser-6xHis-Gly-Ser-

229 235 234

25.6 26.0 25.2

6.11 6.27 6.57

134

¨ GER SOMMER AND BO

FIG. 3. Native molecular weight determination of recombinant corn GST isoforms performed on gel filtration Superose 12 column. Molecular weights (MW) were calculated from linear fit (——) of protein standards (●; ribonuclease A (MW 13.7 kDa), chymotrypsinogen A (MW 25 kDa), ovalbumin (MW 43 kDa), bovine serum albumin (MW 67 kDa), alcohol dehydrogenase (MW 150 kDa), b-amylase (MW 200 kDa), and blue dextran (MW 2,000 kDa)).

active in isomerization of the thiadiazolidine. No isomerization catalyzed by natural GST IV and GST III have been determined. With the same assay conditions these three substrates were also used previously to determine activities of natural corn GST I (conversion of CDNB: 4894 mmol/h 3 mg; metazachlor: 0.07 mmol/h 3 mg; thiadiazolidine BW78: 0.33 mmol/h 3 mg) and GST II (CDNB, 2642 mmol/ h 3 mg; metazachlor, 5.40 mmol/h 3 mg; thiadiazolidine BW78, 13.20 mmol/h 3 mg) (9). Apart from minor variations no significant difference in substrate specificities between recombinant and natural corn GST was detectable. Activities of corn GSTs toward CDNB available in literature, measured in slightly modified assays, are comparable with values in this study (natural GST I (18, 25); heterologous GST I (38); natural GST II (18, 25); natural GST IV was not detected by (22); heterologous GST III (38)). Besides standard GST activities, two substrate analogues to a,b-unsaturated aldehydes were also used as GST substrates (Table 2). Endogenous, toxic a,b-unsaturated aldehydes are produced as a result of free-radical-induced lipid

peroxidation and react with cellular constituents including DNA (32). Crotonaldehyde is a substrate analogue to naturally occurring alkenals formed during the oxidation of fatty acids and nucleic acids (32). Ethacrynic acid (31), an a,bunsaturated ketone, is described as a substrate analogue and an inhibitor of GSTs (39). All four recombinant GST isoforms react with crotonaldehyde with best activity by GST IV. GST isoforms including subunit GST27 had the highest and GST III the lowest activity. No data for the corresponding natural GST isoenzymes are available in literature. Each recombinant GST isoenzyme also showed activity toward ethacrynic acid. Recombinant GST I and GST III showed similar activities, GST II exhibited more than these, and GST IV gave the highest conversion rate. Obviously, the activity for this substrate is due to GST isoforms having a GST27 subunit. Activities available in literature (25) for natural GST I and GST II were higher than in this study. Ethacrynic acid was the better substrate for GSTs than crotonaldehyde (based on the maximum rate (Vmax) to express specific activity).

135

RECOMBINANT GST ISOFORMS TABLE 2 Substrate Specificities of Purified Recombinant Corn GST Isoforms Specific activity [mmol/mg protein 3 h] Substrate

GST I

Thiadiazolidine BW78 CDNB [14C]Metazachlor Crotonaldehyde Ethacrynic acid

6 6 6 6 6

a b

1.05 5120 0.44 1.74 22.74

GST II a

0.15 80a 0.05a 0.11b 1.92b

9.18 3110 5.77 1.86 29.28

6 6 6 6 6

0.45 230 0.83 0.19 0.78

GST IV 11.91 120 4.37 3.46 37.32

6 6 6 6 6

0.25 20 0.65 0.23 0.96

GST III 0.05 530 0.34 1.13 18.96

6 6 6 6 6

0.01 10 0.02 0.23 0.48

Values represent the mean 6 standard deviation (SD; n 5 3). Values represent the mean 6 standard deviation (SD; n 5 5).

Apparently, the additional N-terminal 6xHistag of recombinant GST isoenzymes has no significant influence on substrate specificities and activities. Accordingly, recombinant expression of GSTs in E. coli is a convenient system to provide sufficient amounts of pure GST isoforms for biochemical studies. Glutathione Peroxidase Activities Recombinant GST isoforms were checked for their ability to deactivate cytotoxic organic hydroperoxides. Endogenous products of oxidative damage initiated by hydroxyl radicals (1) were used measuring activities toward a natural occurring unsaturated fatty acid hydroperoxide (linolenic acid hydroperoxide) and two model substrates (cumene hydroperoxide and tert.butyl hydroperoxide). Glutathione peroxidase activities of recombinant GST isoenzymes are shown in Table 3. GST II and GST IV deactivated linolenic acid

hydroperoxide with similar high rates and GST I or GST III with very low rates. Tert.-butyl hydroperoxide was not a substrate for recombinant GST I and GST III. Activity of GST IV was twice as high as of GST II. No data for the corresponding natural GSTs using these two substrates are available in literature. Recombinant GST I showed nearly no activity and GST III low activity toward cumene hydroperoxide. Rate of conversion with GST IV was twice as high as with GST II. Data for natural GSTs in literature (25) showed no activity of GST I and activity of GST II (GST IV and GST III were not determined). We found that linolenic acid hydroperoxide and cumene hydroperoxide were better GST-substrates than tert.-butyl hydroperoxide. GST isoforms including the safener-inducible GST27 subunit showed greater catalytic efficiency in deactivation of organic hydroperoxides than constitutive GST isoforms I and III.

TABLE 3 Glutathione Peroxidase Activities of Recombinant Corn GST Isoforms Specific activity [unitsa/mg protein] Substrate Linolenic acid hydroperoxide tert.-Butyl hydroperoxide Cumene hydroperoxide a b

GST I 11.2 6 0.3 0.02 6 0.01 2.9 6 0.1 b

GST II

GST IV

GST III

547.5 6 18.2 1.81 6 0.06 222.8 6 7.1

553.8 6 31.2 3.80 6 0.08 496.3 6 28.4

16.7 6 0.4 0.04 6 0.02 9.4 6 0.2

1 unit is defined as: 0.868 3 [NADPH oxidized] 3 h21 3 [GSH at start]21 according to ref. (34). Values represent the mean 6 standard deviation (SD; n 5 3).

136

¨ GER SOMMER AND BO

Discussion of Specific Activities Crystal structures of GSTs have revealed that there are kinetically independent binding sites (2, 38) in each GST subunit. So the average of activities of GST I and GST IV for each substrate should be expected when heterodimeric GST II is studied. Surprisingly, specific activities of GST II shown in Tables 2 and 3 do not follow this assumption with every substrate. While activity toward CDNB, ethacrynic acid, tert.-butyl hydroperoxide and cumene hydroperoxide seems to be additional, crotonaldehyde is converted with only half of the rate expected. In contrast, thiadiazolidine BW78 reacts almost 30% faster than expected, metazachlor and linolenic acid hydroperoxide even double the anticipated activity. Probably, conformational shifts occur induced by interaction of subunits. This suggestion is supported by the observations of coexpression of subunits GST29 and GST27 in E. coli. Although a mixture of GST I, GST II, and GST IV was expected, almost no GST I was found. Obviously, formation of GST II is energetically favored under these conditions and remaining GST27 forms homodimer GST IV. Corn GSTs and Oxidative Stress GSTs obviously play a role in oxidative stress tolerance (30). As suggested (25) GSTs should have a further function in protecting plants from injury by cytotoxic agents formed during oxidative stress. When fatty acids are peroxidized, cell components like enzymes, pigments, and membranes may be damaged by free radicals derived from peroxidation process (40). A new corn isoform Zm GST V-V (17) reportedly conjugates alkenal derivates and deactivates cumene hydroperoxide with low rates, but not fatty acid hydroperoxide. GST-mediated glutathione peroxidase activity was also reported for, e.g., pea (41), or Arabidopsis thaliana (36) GSTs. The role of GSTs is broader than anticipated some years ago. First, they were thought to be instrumental as detoxifying enzymes catalyzing conjugation of herbicides and other phytotoxic

compounds. Second, the catalytic isomerization of thiadiazolidine herbicides to their more effective isomeric peroxidizing triazolidine derivatives has to be added as a major activity exhibited by certain GST isoforms (9). Third, the peroxidase activities have to be considered. Surprisingly both bioactivation of thiadiazolidines and deactivation of toxic alkenals or organic hydroperoxides resulting from oxidative stress as reported in this study is catalyzed with high rates by GST isoforms including subunit GST27, which is the major inducible GST subunit in corn (based on both mRNA and protein level) (12, 21). ACKNOWLEDGMENTS We are grateful to Professor Dr. C.-P. Tu, Pennsylvania State University, University Park, PA, for providing plasmid pGTC20 and Professor Dr. I. Jepson, ZENECA Seeds, Bracknell, UK, for plasmids pIJ12 and pIJ21. Due thanks are expressed to K. Wakabayashi and Y. Sato and Department of Agricultural Chemistry, Tamagawa University, Tokyo, Japan, for a gift of thiadiazolidine BW78 and triazolidine BW85.

REFERENCES 1. K. A. Marrs, The functions and regulation of glutathione S-transferases in plants, Annu. Rev. Plant Physiol. Plant Mol. Biol. 47, 127 (1996). 2. T. Neuefeind, R. Huber, P. Reinemer, J. Kna¨blein, L. Prade, K. Mann and B. Bieseler, Cloning, sequencing, crystallization and X-ray structure of glutathione Stransferase-III from Zea Mays var. Mutin: A leading enzyme in detoxification of maize herbicides, J. Mol. Biol. 274, 577 (1997). 3. R. Edwards and D. J. Cole, Glutathione transferases in wheat (Triticum) species with activity toward fenoxaprop-ethyl and other herbicides, Pestic. Biochem. Physiol. 54, 96 (1996). 4. D. E. Riechers, K. Yang, G. P. Irzyk, S. S. Jones and E. P. Fuerst, Variability of glutathione S-transferase levels and dimethenamid tolerance in safener-treated wheat and wheat relatives, Pestic. Biochem. Physiol. 56, 88 (1996). 5. D. J. Cole, Detoxification and activation of agrochemicals in plants, Pestic. Sci. 42, 209 (1994). 6. K. P. Timmerman, Molecular characterization of corn glutathione S-transferase isoenzymes involved in herbicide detoxification, Physiol. Plant. 77, 465 (1989). 7. O. C. Kno¨rzer and P. Bo¨ger, Antagonizing Peroxidizing Herbicides, in “Peroxidizing Herbicides” (P. Bo¨ger and K. Wakabayashi, Eds.), Springer, in press.

RECOMBINANT GST ISOFORMS 8. P. Bo¨ger and K. Wakabayashi, Peroxidizing herbicides (I): Mechanism of action, Z. Naturforsch. 50 c, 159 (1995). 9. B. Nicolaus, Y. Sato, K. Wakabayashi, and P. Bo¨ger, Isomerization of peroxidizing thiadiazolidine herbicides is catalyzed by glutathione S-transferase, Z. Naturforsch. 51 c, 342 (1996). 10. F. Droog, Plant glutathione S-transferase, a tale of theta and tau, J. Plant Growth Regul. 16, 95 (1997). 11. D. P. Dixon, I. Cummins, D. J. Cole and R. Edwards, Glutathione-mediated detoxification systems in plants, Current Opinion in Plant Biol. 1, 258 (1998). 12. I. Jepson, V. J. Lay, D. C. Holt, S. W. J. Bright and A. J. Greenland, Cloning and characterization of maize herbicide safener-induced cDNAs encoding subunits of glutathione S-transferase isoforms I, II and IV, Plant Mol. Biol. 26, 1855 (1994). 13. D. M. Shah, C. M. Hironake, R. C. Wiegand, E. I. Harding, G. G. Krivi and C. Tiemeier, Structural analysis of a maize gene coding for glutathione S-transferase involved in herbicide detoxification, Plant Mol. Biol. 6, 203 (1986). 14. G. P. Irzyk, S. Potter, E. Ward and E. P. Fuerst, A cDNA clone encoding the 27-kilodalton subunits of glutathione S-transferase IV from Zea mays, Plant Physiol. 107, 311 (1995). 15. G. Grove, R. P. Zarlengo, K. P. Timmerman, N. Li, M. F. Tam and C-P. D. Tu, Characterization and heterospecific expression of cDNA clones of genes in the maize GSH S-transferase multigene family, Nucleic Acids Res. 16, 425 (1988). 16. R. E. Moore, M. S. Davies, K. M. O’Connell, E. I. Harding, R. C. Wiegand and D. C. Tiemeier, Cloning and expression of a cDNA encoding a maize glutathione S-transferase in E. coli, Nucleic Acids Res. 14, 7227 (1986). 17. D. P. Dixon, D. J. Cole and R. Edwards, Purification, regulation and cloning of a glutathione transferase (GST) from maize resembling the auxin-inducible typeIII GSTs, Plant Mol. Biol. 36, 75 (1998). 18. T. J. Mozer, D. C. Tiemeier and E. G. Jaworski, Purification and characterization of corn glutathione S-transferase, Biochemistry 22, 1068 (1983). 19. T. Iida, S. Senoo, Y. Sato, B. Nicolaus, K. Wakabayashi ¨ and P. Boger, Isomerization and peroxidizing phytotoxicity of thiadiazolidine-thione compounds, Z. Naturforsch. 50 c, 186 (1995). 20. Y. Sato, P. Bo¨ger and K. Wakabayashi, The enzymatic activation of peroxidizing cyclicisoimide: A new function of glutathione S-transferase and glutathione, J. Pestic. Sci. 22, 33 (1997). 21. D. C. Holt, V. J. Lay, E. D. Clarke, A. Dinsmore, I. Jepson, S. W. J. Bright and A. J. Greenland, Characterization of the safener-induced glutathione S-transferase isoform II from maize, Planta 196, 295 (1995).

137

22. G. P. Irzyk and E. P. Fuerst, Purification and characterization of a glutathione S-transferase from benoxacortreated maize (Zea mays), Plant Physiol. 102, 803 (1993). 23. K. A. Marrs and V. Walbot, Expression and RNA splicing of the maize glutathione S-transferase Bronze2 gene is regulated by cadmium and other stresses, Plant Physiol. 113, 93 (1997). 24. K. M. O’Connell, E. J. Breaux and R. T. Fraley, Different rates of metabolism of two chloroacetanilide herbicides in pioneer 3320 corn, Plant Physiol. 86, 359 (1988). 25. D. P. Dixon, D. J. Cole and R. Edwards, Characterisation of multiple glutathione transferases containing the GST I subunit with activities toward herbicide substrates in maize (Zea mays), Pestic. Sci. 50, 72 (1997). 26. K. A. Marrs, M. R. Alfenito, A. M. Lloyd and V. Walbot, A glutathione S-transferase involved in vacuolar transfer encoded by the maize gene Bronze-2, Nature 375, 397 (1995). 27. V. Dean and T. P. Devarenne, Peroxidase-mediated conjugation of glutathione to unsaturated phenylpropanoids. Evidence against glutathione S-transferase involvement, Physiol. Plant. 99, 271 (1997). 28. J. Sambrook, E. F. Fritsch and T. Maniatis, Molecular Cloning: A Laboratory Manual, 2nd ed. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York (1989). 29. M. M. Bradford, A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding, Anal. Biochem. 72, 248 (1976). 30. O. C. Kno¨rzer, J. Durner and P. Bo¨ger, Alterations in the antioxidative system of suspension-cultured soybean cells (Glycine max) induced by oxidative stress, Physiol. Plant. 97, 388 (1996). 31. W. H. Habig and W. B. Jakoby, Assays for differentiation of glutathione S-transferases, Methods Enzymol. 77, 398 (1981). ˚ . Engstro¨m, J. W. Kozarich 32. K. Berhane, M. Widersten, A and B. Mannervik, Detoxification of base propenals and other a,b-unsaturated aldehyde products of radical reactions and lipid peroxidation by human glutathione transferases, Proc. Natl. Acad. Sci. USA 91, 1480 (1994). 33. L. Flohe´ and W. A. Gu¨nzler, Assays of glutathione peroxidase, Methods Enzymol. 105, 114 (1994). 34. A. Wendel, Glutathione peroxidase, Methods Enzymol. 77, 325 (1981). 35. T. W. Simmons, I. S. Jamall and R. A. Lockshin, Selenium-independent glutathione peroxidase activity associated with glutathione S-transferase from the housefly, Musca domestica, Comp. Biochem. Physiol. 94 b, 323 (1989). 36. D. Bartling, R. Radzio, U. Steiner and E. W. Weiler, A glutathione S-transferase with glutathione peroxidase activity from Arabidopsis thaliana: Molecular cloning

138

¨ GER SOMMER AND BO

and functional characterization, Eur. J. Biochem. 216, 579 (1993). 37. G. Graff, L. A. Anderson and L. W. Jaques, Preparation and purification of soybean lipoxygenase-derived unsaturated hydroperoxy and hydroxy fatty acids and dertermination of molar absorptivities of hydroxy fatty acids, Anal. Biochem. 188, 38 (1990). 38. T. Neuefeind, R. Huber, H. Dasenbrock, L. Prade and B. Bieseler, Crystal structure of herbicide-detoxifying maize glutathione S-transferase-I in complex with lactoglutathione: Evidence for an induced-fit mechanism, J. Mol. Biol. 274, 446 (1997).

39. F. N. J. Droog, P. J. J. Hooykaas, and E. J. van der Zaal, 2,4-Dichlorophenoxyacetic acid and related chlorinated compounds inhibit two auxin-regulated type-III tobacco glutathione S-transferases, Plant Physiol. 107, 1139 (1995). 40. K. J. Kunert, C. Homrighausen, H. Bo¨hme and P. Bo¨ger, Oxyfluorfen and lipid peroxidation: Protein damage as a phytotoxic consequence, Weed Sci. 33, 766 (1985). 41. R. Edwards, Characterisation of glutathione transferases and glutathione peroxidases in pea (Pisum sativum), Physiol. Plant. 98, 594 (1996).